Background: Disability resulting from stroke and traumatic brain injury represent the main causes of long-term complications in adults. There are no universally accepted treatments available to treat these conditions and the financial, personal, familial and social cost of these disabilities cannot be underestimated. Preliminary data from different laboratories have shown that it is possible to modulate plastic processes in the lesioned brain via pharmacological, or brain and somatosensory stimulation techniques. The purpose of this project is to identify mechanisms of stroke motor disability and characterize the most promising techniques to improve cortical plasticity in these patients to enhance functional recovery. Findings this year: Over the past year, we continued a successful international collaboration investigation the use of brain-computer interfaces (BCIs) as novel tools for therapeutic intervention following stroke. The work involved a randomized trial evaluating the long-term effects (up to 6 months) of combined physical therapy and BCI training on functional recovery of severe motor deficits in chronic stroke patients. At the outset of the trial, patients participated in approximately 18 therapy sessions on average, over an 8-week period. Patients were randomly assigned to one of two groups. The first group trained to modulate specific patterns of brain activity called EEG sensorimotor rhythms, which are known to be linked to motor function in healthy individuals. A custom software application was used to analyze the EEG data in real-time. Detected increases in the EEG sensorimotor rhythm were linked to opening of the paralyzed hand via an exoskeleton device, while decreases caused closing of the hand. For a second group of patients, there was no contingency between EEG sensorimotor rhythm modulation and action of the exoskeleton. The paralyzed hand was manipulated by the exoskeleton on a randomized schedule. Greater functional gains, as assessed with the Fugyl Meyer clinical rating scale, were observed in the group receiving contingent BCI training. A total of 30 patients returned at 6 months for a follow-up Fugyl-Meyer assessment. At 6 months, the group receiving real BCI training maintained significantly greater functional gains compared with the sham group. In summary, this work was an important first demonstration that BCI-based rehabilitation promotes long-lasting improvements in motor function of chronic stroke patients, even with severe paresis, and represents a promising strategy in severe stroke neurorehabilitation.